In 1922, Carl Jung began building a stone tower on the upper shore of Lake Zurich, in a village called Bollingen. It had no electricity. No running water initially. No telephone. Jung, who was at the time one of the most sought-after psychologists in Europe, designed the structure not as a home but as a cognitive environment. He called it his "confession in stone." He would arrive by boat, chop his own wood, cook over a fire, and work in a small study with thick walls and a heavy door that closed out every sound except the wind off the lake.
Over the next three decades, Jung produced the major works that would define analytical psychology. The theories of archetypes, the collective unconscious, and psychological types were developed largely in that tower. His correspondence from Bollingen describes a state that sounds unmistakably like what modern neuroscience would recognize as sustained prefrontal engagement: "In my retiring room I am by myself. I keep the key with me all the time; no one else is allowed in there except with my permission... I have done my best work there."
Jung wasn't indulging eccentricity. He was engineering the preconditions for a brain state that Cal Newport, nearly a century later, would call deep work: "Professional activities performed in a state of distraction-free concentration that push your cognitive capabilities to their limit." Newport's 2016 book made the case that deep work is becoming simultaneously more valuable and more rare, creating an asymmetric advantage for anyone who can actually do it.
The neuroscience of deep work shows why Newport's core argument is even more right than his book suggests, and it adds several findings that change how you should practice it. Your brain doesn't just "focus better" without distractions. It enters a qualitatively different mode of operation when sustained attention reaches a threshold, one involving prefrontal network reconfiguration, default mode suppression, and a neurochemical environment that makes the difference between shallow output and work that compounds. What follows is what happens inside the skull during deep work, what the latest research adds to Newport's framework, and why the entrepreneurial obsession with availability may be the most expensive habit in modern business.
Darwin, for context, had his own version of the Bollingen Tower. His daily routine at Down House involved two 90-minute work sessions (morning and afternoon), separated by a walk on his "Sandwalk" thinking path, correspondence, and rest. He produced On the Origin of Species, arguably the most consequential scientific work of the 19th century, on roughly three hours of deep work per day.
What Actually Happens in Your Brain During Deep Work?
The popular understanding of focus is that it's a single dial you turn up. More focus means more of the same brain activity, just intensified. The neuroscience says something more interesting.
When you begin a cognitively demanding task, the dorsolateral prefrontal cortex (dlPFC) activates. This is the executive control center, the region responsible for maintaining task-relevant information in working memory, suppressing irrelevant inputs, and coordinating the cognitive operations required for complex reasoning. The dlPFC is what Newport is implicitly talking about when he describes deep work as pushing your cognitive capabilities to their limit. It is the most metabolically expensive region of the brain, consuming glucose and oxygen at rates that the brain cannot sustain indefinitely.
But the dlPFC doesn't work alone. Within roughly fifteen to twenty minutes of uninterrupted focus on a single task, functional neuroimaging studies show a shift in network dynamics. The task-positive network (TPN), which includes the dlPFC, the posterior parietal cortex, and the premotor areas, becomes increasingly synchronized. Simultaneously, the default mode network (DMN), the constellation of brain regions active during mind-wandering, self-referential thought, and social cognition, becomes increasingly suppressed.
This TPN-DMN anticorrelation is the neural signature of deep work. The two networks exist in a competitive relationship. When one is active, the other quiets. During shallow work, the brain toggles between them rapidly: you write a sentence, check your email, think about what to have for lunch, return to the sentence. Each toggle is a network switch, and each switch carries a cost.
Neuroscientist Gloria Mark at UC Irvine has quantified that cost. Her research found that after an interruption, it takes an average of 23 minutes and 15 seconds to return to the original task at the same level of engagement. That number isn't a soft measure of "feeling focused." It reflects the time required for the TPN to re-establish dominance over the DMN and rebuild the working memory representations that were disrupted by the switch. Every notification, every "quick question," every glance at a phone is not a momentary distraction. It's a 23-minute cognitive reset.
The implications for entrepreneurs are staggering. A founder who checks email or Slack four times during a two-hour work block has not done two hours of deep work minus four brief interruptions. They have done, at best, four fragments of shallow work separated by incomplete network transitions. The brain never reached the TPN-dominant state where complex reasoning, pattern recognition, and novel problem-solving become possible. They sat at the desk for two hours and produced what could have been generated in twenty minutes of genuine deep work.
The Myelin Connection: Why Deep Work Literally Builds Your Brain
Newport's book makes the economic argument for deep work: it's valuable and rare, therefore it commands a premium. The neuroscience adds a biological argument that's even more compelling.
Deep work builds myelin.
Myelin is the fatty insulation that wraps around neural circuits, and it is the physical infrastructure of skill. When a neural pathway fires repeatedly during sustained, focused practice, oligodendrocyte cells respond by layering additional myelin around the active fibers. Each layer of myelin increases the signal speed along that pathway by up to one hundred times and reduces signal loss by a factor that makes the circuit functionally different from an unmyelinated one. The myelinated circuit doesn't just work faster. It works with a precision and reliability that the unmyelinated version cannot approach.
Neurologist George Bartzokis at UCLA spent his career studying myelin development and proposed that the rise and decline of cognitive ability across the human lifespan maps directly to myelin growth and degradation. Children's brains myelinate rapidly, peaking in the late 30s to early 40s, then gradually declining. The circuits that get myelinated are the ones that get used intensely and repeatedly. This is the biological basis of deliberate practice: the reason a pianist who practices four hours of focused scales develops faster than one who plays casually for eight hours is that the focused practice produces more intense, more targeted myelination.
Deep work is the cognitive equivalent of deliberate practice for entrepreneurial skills. When a founder spends ninety uninterrupted minutes building a financial model, the neural circuits involved in quantitative reasoning, assumption testing, and scenario analysis get myelinated. When a founder spends ninety minutes context-switching between the financial model, email, Slack, and a product review, the circuits that get myelinated are the ones for context-switching. The brain builds the infrastructure for whatever you actually do, not whatever you intend to do.
This has a compounding effect that Newport gestures at but doesn't fully articulate. Deep work doesn't just produce better output in the moment. It builds neural circuits that make future deep work more efficient, more precise, and more capable. The founder who protects three hours of daily deep work for a year has literally different brain hardware than the one who spent the same year in a state of constant reactivity. The circuits are faster, the signal-to-noise ratio is higher, and the cognitive ceiling is elevated. The two founders may have started with identical IQ scores. After a year of divergent practice, they are operating with materially different cognitive infrastructure.
What Cal Newport Gets Right, and What the Latest Research Adds
Newport's framework identifies four "depth philosophies": monastic (eliminate all shallow work), bimodal (alternate between deep and shallow periods), rhythmic (schedule fixed deep work blocks daily), and journalistic (drop into deep work whenever a window opens). The neuroscience supports some of these more strongly than others.
The rhythmic philosophy, where you schedule deep work at the same time every day, aligns most directly with what the neuroscience predicts should work best for most people. The reason is circadian regulation of prefrontal function. The dlPFC's metabolic resources are highest in the morning, approximately two to four hours after waking, when cortisol and glucose availability peak. This isn't a preference. It's a biochemical reality. Functional MRI studies show greater dlPFC activation and better sustained attention performance during the circadian peak compared to the trough.
A founder who schedules deep work from 8 to 10:30 AM every day is doing two things simultaneously: leveraging the circadian peak in prefrontal resources and building a conditioned association between a time window and a cognitive state. The brain's habit circuitry, mediated by the basal ganglia, learns to anticipate the deep work state, reducing the transition cost from shallow to deep mode over time. After several weeks, the 23-minute recovery that Gloria Mark measured begins to shrink for these scheduled sessions, because the brain has learned that this time period means TPN dominance, and the network transition initiates automatically.
What the latest research adds to Newport's framework involves the role of the default mode network during rest periods between deep work sessions. Newport treats rest as recovery, which it is. But neuroscience research published since his book reveals that the DMN isn't just resting during these periods. It's doing something actively productive.
Mary Helen Immordino-Yang at USC has shown that the default mode network is where the brain consolidates learning, integrates new information with existing knowledge structures, and generates the kind of insight that feels like it "came from nowhere." The DMN is not the enemy of deep work. It's the complement. During deep work, the TPN processes information intensely. During rest, the DMN integrates that information into long-term memory and connects it with other stored knowledge. The insight you have in the shower, the solution that arrives during a walk, the creative connection that appears the morning after a hard day of work, these are DMN products, and they depend on the deep work session that loaded the raw material.
This means the deep work protocol should include not just focused work sessions but also deliberate, undistracted rest. The founder who finishes a deep work block and immediately picks up their phone to scroll social media is disrupting the DMN integration process. The phone re-engages attentional networks and pulls the brain out of the integrative mode before consolidation is complete. The walk without a podcast, the lunch without a screen, the commute without a call, these aren't wasted time. They're the second half of the deep work cycle.
The flow state literature connects here. Flow, as described by Csikszentmihalyi and explored through Arne Dietrich's transient hypofrontality hypothesis, involves the temporary deactivation of self-monitoring circuits while sensorimotor and creative circuits run at full capacity. Deep work and flow are not identical states, but deep work is the prerequisite condition that makes flow possible. You cannot enter flow while context-switching. The TPN dominance of deep work is the runway that flow's prefrontal quieting launches from.
Why Is Deep Work Disappearing, and What's the Cost?
Newport argues that deep work is becoming rare because the modern knowledge economy rewards visible busyness over invisible depth. The neuroscience adds a mechanism that makes the problem self-reinforcing.
Every time you respond to a notification, check a message, or switch contexts, the brain's reward system delivers a small hit of dopamine. The notification is a variable-ratio reinforcement schedule, the same pattern that makes slot machines addictive. You don't know which email will be important, which Slack message will be urgent, which notification will carry good news. The uncertainty is what makes the behavior compulsive, because the dopamine system responds more strongly to unpredictable rewards than to predictable ones.
Over time, this pattern reshapes the brain's baseline arousal state. The prefrontal cortex becomes conditioned to expect frequent stimulation. Sustained attention on a single task, without any novel input, begins to feel uncomfortable. The brain interprets the absence of stimulation as a signal that something is wrong, and the anterior cingulate cortex generates a restless urge to check, to switch, to seek. This is not laziness. It is a trained neural response, and it becomes stronger with every repetition.
Nicholas Carr explored this territory in The Shallows, arguing that the internet is literally rewiring the brain for shallow processing. The neuroscience supports his core claim. The brain's capacity for sustained attention is use-dependent. Circuits that are used intensely get myelinated and strengthened. Circuits that are abandoned get pruned. A decade of constant context-switching doesn't just create a habit. It restructures the neural architecture for sustained attention, making deep work harder at the hardware level.
For founders, this creates a vicious cycle. The startup environment demands constant availability. Slack channels, investor updates, customer escalations, team questions. The founder who is always available trains their brain for shallow processing. The shallow processing makes deep work increasingly uncomfortable. The discomfort leads to avoidance. The avoidance means the company's hardest problems, the ones that require sustained, deep cognition, never get the brain state they require. Strategy degrades. Product thinking degrades. The creative process that produced the original insight gets replaced by reactive management that produces incremental tweaks. The founder who built something original becomes the operator who can't think past next week.
Try This: The Deep Work Architecture Protocol
A six-step system for building a deep work practice that aligns with the neuroscience of sustained attention, myelination, and network dynamics.
-
Identify your circadian peak. For one week, rate your subjective focus quality every hour on a 1-10 scale. Most people will find a two-to-three-hour window, typically mid-morning, where the scores are consistently highest. This is your deep work window. Protect it the way you would protect a meeting with your most important investor, because the cognitive output it produces is worth more than any meeting.
-
Design a transition ritual. The 23-minute recovery cost from interruption drops significantly when the brain has a consistent cue that deep work is beginning. Newport calls this a "shutdown ritual" for ending work; you need a startup ritual for beginning it. Close every tab. Put your phone in another room, not on silent, not face-down, in another room. Open only the document or tool required for the session. Pour a specific beverage. The ritual becomes a conditioned stimulus that triggers the TPN-dominance shift, and over weeks the transition time compresses.
-
Set a 90-minute ceiling. The dlPFC cannot sustain full engagement indefinitely. Ultradian rhythm research suggests that the brain cycles through periods of high and low arousal approximately every 90 to 120 minutes. Pushing past the natural cycle doesn't produce more deep work. It produces degraded work that feels effortful but lacks the network coherence that characterizes genuine depth. One 90-minute block of real deep work produces more than three hours of forced attention past the point of diminishing returns.
-
Protect the DMN integration window. After your deep work block, take twenty to thirty minutes of undistracted rest. No phone. No email. No podcasts. Walk, sit, stare out a window. This is not procrastination. It is the second half of the deep work cycle, the period when the default mode network integrates what the task-positive network just processed. The solutions and connections that emerge during this window are direct products of the deep work session. Disrupting this window with stimulation is like pulling bread from the oven ten minutes early.
-
Track depth, not hours. Create a simple log: date, deep work start time, end time, and a 1-10 quality rating. At the end of each week, count total deep work minutes. Most founders discover they're averaging less than two hours per day of genuine depth, often less than one. The goal is not to maximize hours. The goal is to ensure that the hours spent in depth are protected, consistent, and high-quality. The morning routine research suggests that optimizing the first ninety minutes of cognition produces outsized returns compared to adding more hours at lower quality.
-
Batch shallow work into compression blocks. Email, Slack, administrative tasks, and low-cognitive-demand communication should be batched into dedicated blocks outside the deep work window. The brain handles shallow work more efficiently in concentrated bursts than in the constant trickle that most founders default to. Set two or three 30-minute blocks for communication. Outside those blocks, every channel is closed. The initial discomfort of being unreachable is real, but it typically resolves within two weeks as teams adapt and the founder discovers that almost nothing is as urgent as it felt.
Carl Jung died in 1961. The Bollingen Tower still stands. Visitors report that the building feels less like a house and more like an environment designed to produce a specific cognitive state. The thick walls block sound. The absence of electricity removes the temptation of artificial stimulation. The view of the lake provides visual rest without engagement. Jung didn't have the vocabulary of network neuroscience, but he understood the principle: the brain's deepest work requires an environment that systematically removes everything that competes for the same neural resources.
You don't need a stone tower. You need ninety minutes, a closed door, a phone in another room, and the willingness to tolerate the discomfort that comes when a brain conditioned for constant stimulation encounters sustained silence. The discomfort is not a sign that something is wrong. It is the transition cost of switching from the network configuration that feels normal to the one that actually produces your best work.
The full neuroscience of sustained attention, prefrontal network dynamics, and the biological mechanisms behind why deep work produces compounding cognitive returns is in Chapter 3 of Wired. If you've ever wondered why your best ideas come in the shower rather than at your desk, the chapter explains the network mechanics and how to engineer them deliberately.
FAQ
What is deep work and what happens in the brain during it? Deep work, as defined by Cal Newport, is professional activity performed in a state of distraction-free concentration that pushes cognitive capabilities to their limit. Neurologically, deep work involves sustained activation of the task-positive network (TPN), including the dorsolateral prefrontal cortex, paired with suppression of the default mode network (DMN). This anticorrelation between networks is the neural signature of depth. It typically takes 15-20 minutes of uninterrupted focus to establish, and research shows it takes an average of 23 minutes and 15 seconds to re-establish after any interruption.
Why does deep work build myelin and why does that matter? When neural circuits fire repeatedly during sustained, focused activity, oligodendrocyte cells wrap additional myelin insulation around the active fibers. Each layer of myelin increases signal speed by up to 100x and reduces signal loss dramatically. This means deep work doesn't just produce better output in the moment; it physically restructures the brain's circuitry to make future work more efficient and precise. The founder who protects daily deep work for a year has literally different neural hardware than one who spent that year in constant reactivity.
How long should a deep work session last? Ultradian rhythm research suggests the brain cycles through periods of high and low arousal approximately every 90-120 minutes. A single 90-minute block of genuine deep work, timed to your circadian peak (typically 2-4 hours after waking), produces more high-quality output than three hours of forced attention past the point of diminishing returns. The dorsolateral prefrontal cortex cannot sustain full metabolic engagement indefinitely, and pushing past the natural cycle produces degraded work that feels effortful but lacks the network coherence of genuine depth.
Is the default mode network bad for productivity? No. The default mode network is essential to the deep work cycle. While the TPN processes information intensely during focused work, the DMN integrates that information into long-term memory, connects it with existing knowledge, and generates insights during rest periods. The "shower insight" and "morning-after solution" are DMN products that depend on the deep work session that loaded the raw material. Disrupting the DMN integration window by immediately checking your phone after a deep work session is like removing bread from the oven before it finishes baking.
How do I start a deep work practice if I'm addicted to constant stimulation? Start with a consistent transition ritual: close every tab, put your phone in another room, and begin with only the required tool open. Set a modest initial target of 60 minutes and build to 90. The discomfort during the first two weeks is the brain's conditioned response to missing its expected stimulation. This response weakens with consistent practice as the basal ganglia learns to anticipate TPN dominance during the scheduled window. Batch all communication into two or three dedicated 30-minute blocks outside your deep work window. Track depth minutes rather than total hours worked.
Works Cited
-
Bartzokis, G. (2004). "Age-Related Myelin Breakdown: A Developmental Model of Cognitive Decline and Alzheimer's Disease." Neurobiology of Aging, 25(1), 5-18. https://doi.org/10.1016/j.neurobiolaging.2003.03.001
-
Carr, N. (2010). The Shallows: What the Internet Is Doing to Our Brains. W. W. Norton.
-
Fox, M. D., Snyder, A. Z., Vincent, J. L., Corbetta, M., Van Essen, D. C., & Raichle, M. E. (2005). "The Human Brain Is Intrinsically Organized into Dynamic, Anticorrelated Functional Networks." Proceedings of the National Academy of Sciences, 102(27), 9673-9678. https://doi.org/10.1073/pnas.0504136102
-
Immordino-Yang, M. H., Christodoulou, J. A., & Singh, V. (2012). "Rest Is Not Idleness: Implications of the Brain's Default Mode for Human Development and Education." Perspectives on Psychological Science, 7(4), 352-364. https://doi.org/10.1177/1745691612447308
-
Mark, G., Gudith, D., & Klocke, U. (2008). "The Cost of Interrupted Work: More Speed and Stress." Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 107-110. https://doi.org/10.1145/1357054.1357072
-
Newport, C. (2016). Deep Work: Rules for Focused Success in a Distracted World. Grand Central Publishing.
-
Coyle, D. (2009). The Talent Code: Greatness Isn't Born. It's Grown. Bantam Books.
-
Jung, C. G. (1963). Memories, Dreams, Reflections. Recorded and edited by Aniela Jaffe. Vintage Books.